CO2 Liquefaction 2

[Alphite]

I realize this is a common engineering problem, but after hours of searching/calculating I could use some advice.

I'd like to liquefy a CO2 stream (assume 99.9% purity). 5-10 T/d, 20 bar, 40C incoming, cooling to -19.5C at 20 bar.

I understand the heat transfer calcs, and it comes out to around 26 kW cooling required. Given the small scale, I was thinking of something like R-143A as the refrigerant.

This is where I'd like some advice. I assume a packaged refrigeration system is best, so are there any recommendations for vendors? I'd also like to work out the numbers on my own for sizing the refrigerant compressor and condensing coil, so a little advice on this end would be helpful as well (even just a book).

I'd then use a shell in tube or U-tube heat exchanger to cool the CO2, and I'm comfortable running these calcs.


Thanks

 

[georgeverghese]

Strange, at 10 000 kg/day, I get 40kW to liquify this stream to -18degC at 20bar abs ( 60kJ/kg to desuperheat, 280kJ/kg for latent heat of condensation from Perry).
R143a selection seems reasonable - compression from about 1.9bar abs to 22bar abs, and water or air cooling to approx 45degC. Condensor would then be operating at about -30degC on the R143a side, which would give you a comfortable delta t of 10degC at the cold end of the HX.
I can see why you are reaching out for help to estimate compression power - Perry has only thermo props for saturated gas/liquid for R143a, not for superheated gas. And I cant find this on the net either. Try getting this from R143a manufacturers. You'd have close to isentropic compression for a recip machine (>95-98 % isentropic eff for recips).
Screw compressors are cheaper and popular, but you could use a recip machine just as well. Suggest a VFD for capacity control for operational reliablity if you're going for a screw compressor.

Found this on the net

http://www.ethermo.us/Show50Vatemp!258.15!1~press!320!4.htm

 

[georgeverghese]

From that website posted earlier, I've abstracted an average value of 1.3 for Cp/Cv for this R143a in the range of compression from 1.9bar abs to 22bar abs.

This gives a compression power of 30kW absorbed for a recip machine option on my compressor performance spreadsheet, a R143a flash liquid fraction of 0.42 at the CO2 condensor/R143 evaporator(assuming no R143a subcooler upstream). R143a recirculating mass rate is 1585kg/hr.
You'd get a slight reduction in compression power with a subcooler. Typically, a feedgas superheater is also included to keep the R143a a little away from dewpoint conditions at the inlet to the compressor.

 

[georgeverghese]

There is more thermo info on Suva HP62 from DuPont on the net - it has no chlorofluoro hydrocarbons either, and it looks like it could work for this service. Interested ?

http://www.pchetz.com/_Uploads/dbsAttachedFiles/h51589_Suva404A_pressure_enthalpy_eng.pdf

 

[Alphite]

Thank you very much for the help!! This definitely gets me further along. I'll try to build out my model more before I bombard you with questions.

And yes, my energy required was off, I was accidently giving the number I had at 5 T/d. It was 21-26 kW at that flowrate, depending on the constants I used, which is close to your 40 kW at 10 T/d.

This is assuming a single stage compressor? A two stage would help limit the max temp, or it doesn't matter with the subcoolers?


Thanks

 

[georgeverghese]

At a Cp/Cv of 1.3, discharge temp is 116degC for a recip compressor, so this is well within limits for a single stage machine. The downstream subcooler would have no effect on this discharge temp. I dont have a high degree of confidence on this Cp/Cv value though, since this is abstracted from that unverified website I posted earlier. Which is why I suggested using this DuPont refrigerant Suva HP62 - the Mollier chart for this appears to have enough info to derive Cp/Cv - there is some thermo and composition info on this HP62 in Perry also.

This looks a lot more reliable for DuPont SuvaHP 62

http://www.allchemi.com/download/tables/R-404A-SI.pdf

Think you'll probably need a separate liquid CO2 accumulator; the subcooler cooling bundle could run through the liquid space in this accumulator.

At 40kW compression duty, you may get offered a twin screw oil flooded machine with slide valve for capacity control based on cost. I'd go for the recip compressor option with suction valve unloader for capacity control on the grounds of long term operational reliability.

Not sure how you'd deal with PSV discharges for pure CO2 service when PSV downstream conditions predict solid CO2 formation, since downstream pressure will be <5bar abs.

 

[pierreick]

Hi,
If you need "reliable data" about refrigerants and other chemicals use this link.

https://webbook.nist.gov/chemistry/fluid/

Good luck
Pierre

 

[pierreick]

Hi,
Attached a document from Art Montemayor.
Good luck
Pierre

 

[georgeverghese]

Error in my post on the 3rd of Dec, compression power is 30kW, not 40kW.

Based on the data abstracted from the link @pierre provided to the NIST, compression power is 26kW for R143a, but compressor discharge temp is 67.5degC for a recip option, with isentropic eff at 100%. I'd tend to go with this result and not my previous result of 116degC. Cp/Cv remains at 1.3 for this compression (1.2 at inlet, 1.4 at exit). Hence we can confirm a single stage recip will do. Discharge temp for an oil flooded screw will be even lower since compression heat is absorbed into lube oil.

Hence the R143a water (or air) cooler / condensor duty must be = heat of compression + heat absorbed for CO2 liquifacton = 40+26 = 66kW (without duty design margin. Condensing temp = 48degC at 22bar abs.

 

[Alphite]

Thank you both, this was incredibly helpful. Saved me a lot of headaches =)